Neuroimaging correlates of prior immunosuppression and cognitive status in HIV.

Type: Poster
Title: Neuroimaging correlates of prior immunosuppression and cognitive status in HIV.
Authors: Fennema-Notestine C, Taylor MJ, Notestine RJ, Wolfson T, Gamst AC, Archibald SL, Theilmann RJ, Heaton RK, Marra C, Grant I, for the CHARTER Group
Date: 03-05-2013
Abstract:Background: Structural MRI and single voxel MR spectroscopy (MRS) studies broadly support reduced neuronal integrity and increased neuroinflammation in HIV. Specific associations between regional metabolite levels, HIV-related variables, and cognitive performance, however, vary across studies, in part due to voxel tissue composition, use of ratios to creatine (CR), and the cohort studied. Using data from the CNS HIV Anti- Retroviral Therapy Effects Research (CHARTER) study, we explored approaches to increase the sensitivity to detect HIV-related effects and identify neuroimaging correlates of HIV-associated neurocognitive impairment (HAND). Methods: MRS concentrations of N-acetylaspartate (NAA), choline, myo-inositol, and CR were quantified using LCModel in frontal white matter (FWM), frontal gray matter (FGM), and basal ganglia (BG) (n=263; 74% on ART). Associations between metabolite concentrations, control variables (e.g., voxel tissue composition, age) and HIV-related factors (plasma HIV RNA, nadir and current CD4) were examined using Spearman's correlation, Mann-Whitney U, and multivariable regression models. Associations between MRS and structural MRI measures and HAND (n=68 with HAND) were explored using stepwise logistic regression. Results: Absolute CR concentration was higher in individuals with detectable plasma HIV RNA (FGM p=.012; FWM p=.008; BG p=.009), therefore, ratios to CR were not used. In multivariable models, effects of age and voxel tissue composition were common; for example, older individuals had lower FGM NAA (t=-2.0, p=.04), and FGM voxels containing more gray matter had higher NAA (t=4.1, p=.0001). Controlling for such factors, NAA was lower in individuals with lower CD4 nadir (FGM t=2.4, p=.02; BG t=2.5, p=.01). Models selected lower FGM NAA (t=-3.2, p=.001) and more abnormal white matter (t=2.1, p=.03) to best predict HAND diagnosis, even only within individuals currently on ART (FGM NAA t=-2.7, p=.007; abnormal white matter t=1.9, p=.06). Conclusions: Poorer neuronal integrity was seen in individuals with greater immunosuppression historically, supporting earlier cognitive findings in CHARTER. HAND was most strongly associated with lower neuronal integrity and more extensive abnormal white matter, which may reflect persistent HIV-related inflammation. Treatment initiated to prevent low CD4 nadir may reduce the extent of brain injury and cognitive impairment in HIV, as we continue to explore the impact of ongoing neuroinflammation.